My RAM finally came today so Ive been able to try some overclocks properly
Been messing around with the GTL's a lot
Hey Ducky Spud, what RAM did you get then?
I found GTL values of 0.700, 0.680 and 0.610 worked best. Mines an old C0 stepping chip though
Ah I see you changed the second CPU-GTL from x67 to x68, I've not tried changing that one yet so that one is interesting to me.
WHat is the effect on your machine from adjusting that second CPU-GTL from x67 to x68 btw? does it resolve a no-boot situation or does it bring stability?
Can I ask if you could do a list of what you changed at what point in bios.
Hey Clairvoyant,
wish I had a list for that but I'm ashamed to admit I am using a combination of
Plodding meets
The Force meets
Chaos Theory heh!
If I was to try and make some sense from it I pretty much start at full bios [Auto] and then slowly but surely start adjusting things one-by-one and testing lots (reboots, primes, IBT etc), the plodding requires patience but makes it really obvious when a setting doesn't i.e the PC is rock-stable then I change one setting and it blows up!
. . . . doesn't take great detective work to figure out what went wrong!
After too much plodding Patience begins to wear thin and you start changing several BIOS paremeters at a time and with some luck you begin making quantum jumps which feels nice . . . however if your not lucky or your theory is unsound everything blows up and your left scratching your head heh!
I'd say the bulk of the voltage controls are where my attention is most of the time, starting from full BIOS [Auto] I change the processor voltage to manual and begin lowering it, notch by notch with Prime Small FFTs and IBT runs inetween to gauge stability, I keep going until the PC fails to POST or load Windows Etc and then slowly increase the vCore until everything is rock stable. The resulting figure gives me a baseline for processor scaling and is essentially a bespoke tested
REAL-VID.
Then one by one I go through the voltage adjustments lowing their values until they are rock bottom. I think the P5Q series can pretty much run a 400MHz-FSB with all the main voltages on minimum (or minimum +1 is using an older BIOS that overvolts say).
I'm not sure when is the best time to manually set GTL-Refs though, they seem fine on [Auto] until you begin climbing high with the FSB (somewhere above 400MHz-FSB).
I will try and make a more useful post to you in a week or two as I am starting a new clock so my second time around may yeild some useful info.
I have a new E0 stepping with a VID of 1.25 (not good but so far testing ok @ 1.18/9 underload)
Nice one!
That sounds the same as the chip I just sold, that was a 1.2500 VID and it clocked well. The low VID chips may require a bit less voltage to run stably but as gurusan has mentioned before the low VID chips seem to run a lot hotter than a high VID chip.
For someone with a high grade water-cooling system this poses no problem but for people who use Air-Cooling (like me!) a Higher VID chip could be preferable as you can get big clocks and still keep the little CPU's well within specs, this 1.1875v VID chip I am testing deffo runs hotter than my old 1.2500v VID chip so count your blessing!
Look forward to everyones results and findings, stabililty screens and pictures of your systems also greatly appreciated!
